All over the world, the oil and gas industries are the most important source of energy. The likelihood of subsea gas and oil leakage is increasing, which can cause threats and harmful impacts on the marine environment and potentially catastrophic events such as fires, explosives, and the loss of structural integrity of subsea infrastructure. Also, Physical models have usually been used to predict sea currents, but they are unstable to disturbances and hence incorrect over long periods of time. Machine learning approaches are more resistant than the physical models that have usually been used to predict sea currents. Also, Physical models have usually been used to predict sea currents. Nonetheless, they are not stable to disturbances and thus are not correct for long periods of time. Machine learning approaches are more resistant than the physical models that have usually been used to predict sea currents. Machine learning approaches are more resistant to change and perturbation. Therefore, the goal of this research is to assess the potential of hazards of the gas plume from subsea pipeline rupture till reaching the sea surface by changing the influence parameters on gas plume to assist petroleum companies in developing risk assessment strategies by assessing and simulating subsea gas release in order to contain the leakage by developing coupling models one for machine learning code which can predict the upcoming water current speed by using Multiple Linear Regression algorithm and hooked it by UDF to a second model which implements Computational Fluid Dynamics (CFD) model to study subsea gas release under current effects. Engebretsen’s Rotvoll experiment data is being used to validate the numerical computational fluid dynamics model. The rising time and fountain height and horizontal migration for gas release are the essential factors to be considered while evaluating the gas dispersion through our study by changing the influencing parameters such as leakage hole sizes, water current speeds, gas velocity, and water depths in the presence of water current in all cases. Also, applied our simulation to real case parameters for one of the Egyptian Petroleum Companies. These findings might aid in evaluating the hazards and response planning in the event of subsea gas leakage.
Modeling and assessment of accidental subsea gas leakage using a coupled computational fluid dynamics and machine learning approaches
2023-08-01
24 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
British Library Conference Proceedings | 2014
|Effective adjoint approaches for computational fluid dynamics
Elsevier | 2019
|